Three-coordinate measuring and positioning system and method for repair welding of drawing type friction plug

Abstract

The invention provides a three-coordinate measuring and positioning system and method for drawing type friction plug repair welding. In the drawing type friction plug repair welding measuring and positioning process, S1, a robot is operated to drive a contact type measuring head to conduct multi-path scanning along the X axis, the Y axis and the Z axis of a storage box, three-dimensional coordinate data are collected, and an initial three-dimensional contour point cloud is generated; s2, preprocessing the initial point cloud, and performing measuring head radius offset compensation and temperature drift compensation to obtain corrected point cloud data; and S3, reconstructing a curved surface by using a non-uniform rational B-spline algorithm, generating a high-precision three-dimensional model of the storage tank, and determining a defect center position and a normal vector. And S4, related data are transmitted to a robot control system through bus communication, and an obstacle avoidance motion path is generated in combination with laser ranging data. And S5, the posture of the robot is adjusted so that the tool can be aligned with the defect, monitoring is conducted through a visual sensor, dynamic secondary compensation is triggered and the path is updated when the deviation exceeds a threshold value, and accurate repair welding is ensured.

Description

Technical Field

[0001] This invention relates to the field of automated measurement and welding technology, and in particular to a coordinate measuring and positioning system and method for pull-out friction plug repair welding. Background Technology

[0002] In aerospace, energy, and many other fields, the manufacture and maintenance of large equipment such as aerospace storage tanks and high-pressure vessels are crucial. During long-term use, various defects inevitably appear on the surface of these devices, especially typical defects such as round holes, which seriously affect the safety and reliability of the equipment.

[0003] Taking aerospace storage tanks as an example, as a critical component of spacecraft, they must withstand enormous pressure in the complex environment of space, and even minor surface defects can lead to serious consequences. Traditional defect measurement methods have low accuracy and are difficult to meet the stringent requirements for equipment manufacturing and maintenance in high-end fields such as aerospace.

[0004] In previous measurement techniques, some methods could only obtain general information about the surface and could not accurately measure key parameters such as the size, location, and shape errors of defects. Moreover, the measurement process is easily affected by environmental factors. For example, temperature changes can cause the materials of the equipment to expand and contract, leading to deviations in the measurement results; errors in the measurement sensors themselves can further reduce the measurement accuracy.

[0005] Furthermore, when processing and analyzing measurement data, limitations in algorithms and models make it difficult to accurately reconstruct the surface model of the equipment, thus failing to provide a reliable basis for subsequent repair work. During the repair operation, the robot's positioning and attitude adjustment lack precise guidance, resulting in inaccurate welding positions and angles, affecting welding quality, and making it difficult to ensure that the equipment's performance is restored to its ideal state.

[0006] Therefore, it is urgent to develop a high-precision, interference-resistant three-coordinate measurement and positioning system and method for pull-out friction plug repair welding. Summary of the Invention

[0007] This application provides a coordinate measuring and positioning system and method for pull-out friction plug repair welding to solve the problems in the background art.

[0008] In a first aspect, the present invention provides a coordinate measuring and positioning system for pull-out friction plug repair welding, comprising: The robot body has a spindle tool holder at its end. A coordinate measuring sensor is mounted on the spindle tool holder via a clamping adapter, and is used to contact the surface of the tank and collect the coordinates of the measurement points. The control subsystem is used to latch the grating signal of the measurement sensor, record the three-dimensional coordinates of the measurement point and generate point cloud data. The data processing module is used to perform surface fitting on the point cloud data, reconstruct the tank surface model, and calculate the center position and normal of the defect; The robot controller is used to adjust the robot's posture according to the center position and normal of the defect so that the tool zero point coincides with the center of the defect and the cutting tool coincides with the normal.

[0009] Furthermore, the clamping adapter includes: The tool holder connection part is detachably connected to the spindle tool holder; The probe mounting part is used to fix the coordinate measuring sensor, and a radius compensation adjustment structure is provided between the probe mounting part and the tool holder connection part to compensate for the influence of the probe radius on the measurement accuracy.

[0010] Secondly, the present invention provides a coordinate measuring and positioning method for welding pull-out friction plugs, applied to the coordinate measuring and positioning system for welding pull-out friction plugs as described above, comprising the following steps: S1. The robot drives the contact probe to perform multi-path scanning along the X, Y, and Z axes of the storage tank, collects the three-dimensional coordinate data of the measurement points, and generates an initial three-dimensional contour point cloud. S2. Preprocess the initial three-dimensional contour point cloud, including contact point offset compensation based on probe radius and temperature drift compensation based on material linear expansion coefficient, to obtain corrected point cloud data. S3. The non-uniform rational B-spline algorithm is used to reconstruct the surface of the corrected point cloud data to generate a high-precision three-dimensional model of the tank's outer surface, and the center position and normal vector of the defect are determined based on the model. S4. The defect center coordinates and normal vector are transmitted to the robot control system via bus communication. Combined with the defect edge data scanned synchronously by the laser rangefinder, the robot obstacle avoidance motion path is generated. S5: Adjust the robot's posture to make the tool zero point coincide with the defect center and the tool axis aligned with the normal vector, and monitor the alignment status of the tool and the defect in real time through a vision sensor; if the deviation exceeds the preset threshold, trigger dynamic secondary compensation and update the robot's motion path.

[0011] Furthermore, the preprocessing in step S2 also includes: The Kalman filter algorithm is used to remove noise from the initial point cloud data; Based on the real-time temperature data of the tank material, the coefficient of linear expansion is dynamically adjusted, and the temperature drift compensation is calculated.

[0012] Furthermore, the generation of the robot's obstacle avoidance path in step S4 includes: The robot's joint motion trajectory is planned based on the A* algorithm to avoid obstacles inside the storage tank; A path curvature constraint is introduced to ensure that the deviation between the tool and the defect normal is less than 0.1 mm during the tool movement. The path parameters are transmitted to the robot controller in real time via bus communication, and the risk of joint angle exceeding the limit is monitored.

[0013] Furthermore, the dynamic secondary compensation in step S5 includes: The real-time relative position data between the tool and the defect is collected by a vision sensor, and the deviation matrix is ​​calculated. The iterative nearest point algorithm is used to match the actual point cloud with the theoretical model to generate the compensated robot target pose. After updating the motion path, repeat steps S1 to S3 until the alignment deviation meets the preset accuracy requirements.

[0014] Furthermore, the specific steps of NURBS surface reconstruction in step S3 are as follows: Extract the feature control points from the corrected point cloud data and assign weight coefficients to each control point; The node vector distribution is adaptively adjusted based on the defect geometry to optimize surface smoothness. The NURBS surface equation is fitted using the least squares method, and the coordinates of the defect center and the normal vector are output.

[0015] Furthermore, it also includes step S6: After the welding is completed, a laser scanner is used to perform three-dimensional morphological inspection of the repaired area and generate a quality assessment report. If residual gaps or deformation exceeding tolerances are detected, mark the defect location and trigger the rework process.

[0016] The above-described technical solution of the present invention has the following advantages: This invention provides a coordinate measuring and positioning system and method for welding pull-out friction plugs. The system utilizes a robot body to drive a coordinate measuring sensor mounted on the spindle tool holder, enabling flexible movement and multi-path scanning along the X, Y, and Z axes of the tank. The coordinate measuring sensor accurately acquires the coordinates of the measurement points, and the control subsystem latches the grating signal and records the three-dimensional coordinates to generate point cloud data. This series of operations ensures comprehensive and accurate acquisition of tank surface information, significantly improving measurement accuracy compared to traditional measurement methods and providing a reliable data foundation for subsequent analysis.

[0017] The data processing module performs surface fitting and model reconstruction on the collected point cloud data, accurately restoring the surface shape of the storage tank. Based on this, it calculates the center position and normal of the defect, providing precise position and direction information for welding repair, making the welding operation more targeted, avoiding blind repair, and improving the accuracy and success rate of repair.

[0018] The robot controller adjusts the robot's posture based on the center position and normal information of the defect, ensuring that the tool's zero point coincides with the defect's center and the cutting tool coincides with the normal. This ensures that during the repair welding process, the welding tool can accurately act on the defect location, and the welding angle meets the requirements, guaranteeing the quality of the repair welding, reducing repair welding defects caused by positioning deviations, and improving the reliability and stability of the repair welding.

[0019] The entire technical solution automates the process from measurement and data processing to robot operation. The close coordination of each component reduces human intervention, shortens measurement and positioning time, improves work efficiency, lowers labor costs and human error, and is more suitable for large-scale production and equipment maintenance scenarios. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the specific embodiments of this application or the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 A flowchart of a coordinate measuring and positioning method for welding a pull-out friction plug, provided in an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the measurement method for circular hole defects on the outer surface of a storage tank, using a circular defect as an example. Detailed Implementation

[0022] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of the invention. However, those skilled in the art will understand that the invention can be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the invention with unnecessary detail.

[0023] It should be understood that, when used in this specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0024] Furthermore, in the description of this invention and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0025] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of the invention include a particular feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized. "A plurality" means "two or more."

[0026] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0027] According to a first aspect of the present invention, a coordinate measuring and positioning system for pull-out friction plug repair welding is provided, which may include: The robot body has a spindle tool holder at its end. A coordinate measuring sensor is mounted on the spindle tool holder via a clamping adapter, and is used to contact the surface of the tank and collect the coordinates of the measurement points. The control subsystem is used to latch the grating signal of the measurement sensor, record the three-dimensional coordinates of the measurement point and generate point cloud data. The data processing module is used to perform surface fitting on the point cloud data, reconstruct the tank surface model, and calculate the center position and normal of the defect; The robot controller is used to adjust the robot's posture according to the center position and normal of the defect so that the tool zero point coincides with the center of the defect and the cutting tool coincides with the normal.

[0028] In this invention, the robot body serves as the motion execution carrier for the entire system, possessing multi-axis motion capabilities and the ability to move flexibly in three-dimensional space. The spindle tool holder at its end plays a crucial role in connection and positioning, providing an interface for the coordinate measuring sensor. The spindle tool holder not only ensures the stability of the coordinate measuring sensor during robot movement but also precisely controls the sensor's position and orientation, enabling it to accurately contact the tank surface for acquiring coordinate measurements.

[0029] The coordinate measuring machine (CMM) sensor is the core component for directly acquiring measurement data from the tank surface. It is mounted on the spindle tool holder via a clamping adapter, a connection method that facilitates installation and disassembly while ensuring sensor stability. The design of the clamping adapter is particularly crucial, featuring a radius compensation adjustment structure to compensate for the influence of the probe radius on measurement accuracy. In actual measurements, because the probe has a certain radius, the measurement point is not the actual contact point on the tank surface. This radius compensation adjustment structure transforms the coordinates of the measurement point into the coordinates of the actual contact point, improving measurement accuracy. When in contact with the tank surface, the CMM sensor can quickly and accurately acquire the coordinate information of the measurement point, providing raw data support for subsequent data analysis.

[0030] The control subsystem is the hub of data acquisition and processing. It is responsible for latching the grating signals from the measurement sensors, which contain the position information of the measurement points. By analyzing and processing these signals, the control subsystem can accurately record the three-dimensional coordinates of the measurement points and integrate a large amount of measurement point coordinate data to generate point cloud data. The point cloud data is a preliminary digital representation of the tank surface shape, providing the foundation for subsequent data processing and model reconstruction. The control subsystem's efficient data processing capabilities and precise signal latching function ensure the accuracy and integrity of the measurement data.

[0031] The data processing module performs in-depth analysis and processing of the point cloud data generated by the control subsystem. It employs a surface fitting algorithm to fit the discrete point cloud data into a continuous surface, thereby reconstructing the tank's surface model. During the reconstruction process, mathematical calculations and algorithm optimization ensure that the model accurately reflects the tank's true shape. Based on the reconstructed surface model, the data processing module further calculates the center position and normal of the defect. Determining the defect's center position and normal is crucial for subsequent welding work, providing key information for the robot's precise positioning.

[0032] Based on the defect center position and normal information calculated by the data processing module, the robot controller adjusts the robot's posture. It controls the movement of each joint to align the tool zero point with the defect center and the cutting tool with the normal. This process requires the robot controller to precisely calculate the motion angles and displacements of each joint to ensure the robot accurately reaches the welding repair position and maintains the correct posture. During the adjustment process, the robot controller monitors the robot's motion status in real time and continuously optimizes the adjustment strategy based on feedback information to achieve high-precision positioning, ensuring the accurate implementation of pull-out friction plug welding.

[0033] In some alternative embodiments, the clamping adapter may include: The tool holder connection part is detachably connected to the spindle tool holder; The probe mounting part is used to fix the coordinate measuring sensor, and a radius compensation adjustment structure is provided between the probe mounting part and the tool holder connection part to compensate for the influence of the probe radius on the measurement accuracy.

[0034] In this embodiment, the tool holder connector serves as the bridge between the clamping adapter and the spindle tool holder at the end of the robot body. It employs a detachable connection, which, from an operational convenience perspective, allows for quick and easy removal or installation of the clamping adapter from the spindle tool holder when replacing the coordinate measuring sensor, maintaining the device, or making adjustments, thus improving work efficiency. From an equipment versatility perspective, this design enables the installation of different types or specifications of coordinate measuring sensors on the same robot's spindle tool holder, provided they conform to the interface standard of the tool holder connector. This enhances the applicability and flexibility of the entire measurement and positioning system and reduces equipment costs.

[0035] The core function of the probe mounting section is to provide a stable and reliable fixed support for the coordinate measuring machine (CMM). During the measurement process, the CMM needs to accurately contact the tank surface and collect the coordinates of the measurement points. The probe mounting section must ensure that the sensor does not shake or shift during operation to guarantee the accuracy of the measurement data. Its design matches the shape and structure of the CMM, and through specialized mechanical structures such as slots and bolt connections, the sensor is firmly fixed in a specific position, enabling it to accurately scan and measure the tank surface under the robot's guidance.

[0036] The probe of a coordinate measuring machine (CMM) sensor has a certain radius. During measurement, the actual coordinates acquired are those of the probe's center, not the coordinates of the true contact point on the tank surface, leading to measurement errors. The radius compensation adjustment structure exists to eliminate this error. This structure achieves probe radius compensation through precise mechanical adjustment or algorithm-based compensation. For mechanical adjustment, fine-tunable mechanical components may be used to adjust the relative position between the probe mounting part and the tool holder connection part according to the probe radius, making the measured point coordinates closer to the true contact point coordinates. For algorithm-based compensation, the probe radius parameter is input into the control subsystem or data processing module, and a specific algorithm is used to correct the acquired measured point coordinates. Regardless of the method, the radius compensation adjustment structure effectively compensates for the influence of the probe radius on measurement accuracy, improving measurement accuracy and providing a more reliable data foundation for subsequent operations such as accurately reconstructing the tank surface model, determining defect locations, and normals.

[0037] This invention also provides a coordinate measuring and positioning method for welding pull-out friction plugs, applied to the coordinate measuring and positioning system for welding pull-out friction plugs described above, such as... Figure 1As shown, the following steps may be included: Step S1: The robot drives the contact probe to perform multi-path scanning along the X, Y, and Z axes of the storage tank, collects the three-dimensional coordinate data of the measurement points, and generates an initial three-dimensional contour point cloud.

[0038] The following section describes the main body and execution components of the measurement operation. The operation is driven by a robot, whose multi-axis linkage motion capability allows it to flexibly control the contact probe mounted on the end-effector tool holder to move in three-dimensional space. As a key component for directly acquiring measurement data, the accuracy and stability of the contact probe are crucial. During the scanning process, the probe directly contacts the surface of the tank, and its built-in sensors detect the position information of the contact point, providing the initial basis for subsequent coordinate data acquisition.

[0039] Storage tanks typically have complex curved surface structures, and a single scanning path cannot fully capture their surface information. However, multi-path scanning can effectively avoid measurement blind spots, ensuring that the acquired data can completely and accurately reflect the true shape and defect characteristics of the storage tank surface. For example, for areas with depressions, protrusions, or irregular shapes that may exist on the storage tank surface, scanning with different paths can acquire data from multiple angles and positions, thereby obtaining richer and more detailed surface information, laying a solid foundation for subsequent accurate data analysis and processing.

[0040] As the robot drives the contact probe along the X, Y, and Z axes of the storage tank, it converts the position of each measurement point in three-dimensional space into an electrical or digital signal. Upon receiving these signals, the control system accurately records the coordinates of the measurement points along the X, Y, and Z axes, based on preset measurement accuracy and a coordinate system. As the probe continues to move across the tank surface, a large amount of three-dimensional coordinate data from numerous measurement points accumulates, gradually revealing the contour information of the tank surface.

[0041] The three-dimensional coordinate data of numerous measurement points are gathered together to form an initial three-dimensional contour point cloud. The point cloud data initially presents the surface shape of the tank in the form of discrete points. By preprocessing the initial three-dimensional contour point cloud and reconstructing the surface, a high-precision three-dimensional model of the tank's outer surface can be generated, the center position and normal vector of the defect can be determined, and key data support can be provided for the robot to adjust its posture for precise welding.

[0042] Step S2: Preprocess the initial three-dimensional contour point cloud, including contact point offset compensation based on probe radius and temperature drift compensation based on material linear expansion coefficient, to obtain corrected point cloud data.

[0043] Considering that the initial 3D contour point cloud data is affected by various factors during the acquisition process, resulting in errors and deviations, these errors will accumulate in subsequent stages such as surface reconstruction and defect localization if preprocessing is not performed, severely affecting the accuracy of measurement and positioning, and consequently impacting the quality of pull-out friction plug repair welding. Therefore, preprocessing of the initial 3D contour point cloud is necessary.

[0044] The probe of a coordinate measuring machine (CMM) sensor has a certain radius. When measuring the surface of a storage tank, the actual coordinate points acquired are not the actual contact points on the tank surface, but rather the coordinates of the probe center. This leads to a deviation in the measurement point position, affecting the accurate judgment of the tank surface contour and defect location. By establishing a mathematical model, the actual offset of the contact point is calculated based on the probe radius and the contact state between the probe and the tank surface during measurement. For example, when measuring circular hole defects, geometric relationships and trigonometric functions are used to transform the probe center coordinates into actual contact point coordinates, thus correcting the measurement point position. This effectively eliminates the influence of the probe radius on measurement accuracy, making the subsequent tank surface model constructed based on the measurement points closer to the true shape, improving the accuracy of defect center location and normal calculation, and providing more precise positional information for welding repair.

[0045] Meanwhile, the ambient temperature of the storage tank changes, and the tank material exhibits thermal expansion and contraction properties. Temperature changes lead to alterations in the tank's dimensions, causing deviations in the coordinates of the measurement points and affecting the accuracy of the measurement results. To address this issue, this step involves acquiring real-time temperature data of the tank material and, combined with the material's coefficient of linear expansion, calculating the dimensional changes of the tank due to temperature variations. For example, using the thermal expansion formula, based on the temperature change and the coefficient of linear expansion, the displacement of the measurement point in various directions is calculated, thereby correcting the coordinates of the measurement point.

[0046] To ensure the accuracy and reliability of measurement data under different temperature conditions, the influence of temperature on the measurement results is considered, so that the measurement results can truly reflect the actual state of the tank and provide stable and accurate data support for subsequent welding repair work. After contact point offset compensation based on probe radius and temperature drift compensation based on material linear expansion coefficient, the errors and deviations in the initial three-dimensional contour point cloud data are effectively corrected, generating more accurate corrected point cloud data. This data can provide a reliable data foundation for subsequent operations such as surface reconstruction using non-uniform rational B-spline algorithm, determining the center position of defects and normal vectors, which helps to improve the accuracy and reliability of the entire coordinate measuring and positioning system for pull-out friction plug welding repair, ensuring the smooth progress and quality of the welding repair work.

[0047] Step S3: Use the non-uniform rational B-spline algorithm to reconstruct the surface of the corrected point cloud data, generate a high-precision three-dimensional model of the tank's outer surface, and determine the center position and normal vector of the defect based on the model.

[0048] The NURBS algorithm accurately describes the shape of curves and surfaces by defining control points, weight coefficients, and node vectors. Its advantages lie in its ability to flexibly represent various shapes, its good fitting effect on complex shapes, and the ability to control the shape and smoothness of surfaces by adjusting parameters, making it suitable for reconstructing complex surfaces such as storage tanks.

[0049] The following section details the surface reconstruction process based on the NURBS algorithm. First, key points representing the shape characteristics of the tank surface are selected from the corrected point cloud data as feature control points. The selection of these points affects the accuracy and effectiveness of the surface reconstruction; they are generally chosen from areas with significant curvature changes or critical shape features. Simultaneously, a weight coefficient is assigned to each control point; the larger the weight, the greater the influence of the control point on the surface shape. By appropriately setting the weight coefficients, the surface shape can be better adjusted to better fit the actual tank surface. Then, based on the characteristics of the tank surface and the geometry of defects, the node vector distribution is adaptively adjusted. The node vectors determine the partitioning of the curve or surface in the parameter space; a reasonable node vector distribution can optimize surface smoothness. For areas with defects on the tank, the node density can be appropriately increased near these areas to more accurately describe the surface shape; in areas with relatively smooth surfaces, the number of nodes can be appropriately reduced to improve computational efficiency.

[0050] The least squares method is used to fit the NURBS surface equations to minimize the sum of squared errors between the reconstructed surface and the corrected point cloud data. By solving the relevant equations, the parameters of the NURBS surface equations are determined, thereby generating a high-precision 3D model of the tank's external surface. This model accurately reflects the actual shape of the tank, including details such as the curvature and convexity of the surface.

[0051] On the generated high-precision 3D model, defect areas are identified and located using specific algorithms. For common circular hole defects, the center position of the defect can be determined based on geometric calculation methods, using the coordinate data of the circle centers measured at multiple cross-sections, through fitting or other mathematical methods. For example, by spatially fitting the coordinates of the center centers of multiple cross-section circles, the fitted center is the center position of the defect.

[0052] The normal vector is perpendicular to the surface location of the defect and is crucial for ensuring the correct contact angle between the tool and the defect surface during welding repair. In the NURBS surface model, the normal vector at the center of the defect is calculated through mathematical operations such as taking partial derivatives of the surface equations. Specifically, based on the local geometric properties of the surface at that point, the direction and magnitude of the normal vector are determined using differential geometry methods.

[0053] Through the above steps, the technical solution of this application uses the NURBS algorithm to achieve high-precision reconstruction of the outer curved surface of the storage tank and accurately determine the center position and normal vector of the defect. This information can provide data support for the subsequent robot to adjust its posture so that the tool zero point coincides with the defect center and the tool is aligned with the normal vector, ensuring that the pull-out friction plug repair welding can be carried out accurately, thereby improving the quality and efficiency of the repair welding and ensuring the safety and reliability of aerospace storage tanks and other equipment.

[0054] Step S4: Transmit the defect center coordinates and normal vector to the robot control system via bus communication, and generate the robot obstacle avoidance motion path by combining the defect edge data scanned synchronously by the laser rangefinder.

[0055] The coordinates of the defect center and its normal vector are crucial information for determining the location and direction of the weld repair. Data from the defect edge scanned by a laser rangefinder reflects the surrounding environment. Transmitting and integrating this data to the robot control system is fundamental to generating a reasonable motion path. Only by combining this data can the robot plan a trajectory in the complex tank environment that avoids obstacles while accurately reaching the weld repair location.

[0056] Bus communication, as an efficient data transmission method, is responsible for transmitting the defect center coordinates and normal vector obtained by the data processing module to the robot control system. In this process, the data is packaged, transmitted and unpacked according to a specific communication protocol to ensure that the information arrives at the robot control system accurately and provides precise data support for subsequent path planning.

[0057] The laser rangefinder simultaneously scans the defect edge, acquiring distance information by emitting a laser beam and measuring the time it takes for the reflected light to pass, thus obtaining the contour data of the defect edge. This data can reflect the shape, size, and relative positional relationship of the defect's surroundings to surrounding objects in real time. For example, if there are structures inside the storage tank, the laser rangefinder can detect the distance between these structures and the defect edge, providing crucial information for robot obstacle avoidance.

[0058] After receiving the data, the robot control system can use the A* algorithm to plan the robot's joint motion trajectory. The A* algorithm is a commonly used heuristic search algorithm that comprehensively considers the distance from the starting point to the current point and the estimated distance from the current point to the target point, enabling it to quickly find an approximately optimal path from the starting position to the target position (i.e., the defect center) in complex environments. During the planning process, the algorithm uses defect edge data acquired by the laser rangefinder and information about the internal structure of the storage tank as obstacle information, avoiding these obstacles to plan the robot's joint motion trajectory.

[0059] To ensure that the deviation between the tool and the defect normal is less than 0.1mm during tool movement, this invention introduces path curvature constraints. This means that when planning the motion path, it is necessary not only to avoid obstacles but also to ensure the smoothness and accuracy of the path, so that the tool can operate according to the preset normal vector direction when approaching the defect. For example, by limiting and adjusting the curvature of each point on the path, the robot's movement becomes more stable and precise, avoiding excessive deviation between the tool and the defect normal due to an overly tortuous path, which would affect the quality of the weld repair.

[0060] After path planning is completed, the path parameters are transmitted to the robot controller in real time via bus communication. Simultaneously, the system continuously monitors the robot's joint angles to prevent them from exceeding limits. Exceeding joint angle limits can prevent the robot from moving properly or even damage the equipment. Once a joint angle is detected to be close to or beyond the limit, the system will promptly adjust the path planning to ensure the safety and reliability of the robot's movement.

[0061] Step S5: Adjust the robot's posture so that the tool zero point coincides with the defect center and the tool axis aligns with the normal vector. Monitor the alignment status of the tool and the defect in real time through a vision sensor. If the deviation exceeds a preset threshold, trigger dynamic secondary compensation and update the robot's motion path.

[0062] In this step, the zero point of the tool is aligned with the center of the defect, and the tool axis is aligned with the normal vector. This is a requirement to ensure the accuracy and quality of the repair welding. Only when this state is achieved can it be ensured that the repair welding material is accurately filled into the defect position during the repair welding process, and that the welding direction is perpendicular to the plane where the defect is located. This ensures that the strength and sealing performance of the repair weld meet the requirements, thereby effectively repairing the surface defects of the storage tank and ensuring the safety and reliability of the storage tank.

[0063] Based on the previously obtained defect center coordinates and normal vector information, the robot controller adjusts the robot's posture by controlling the movement of each joint. Using inverse kinematics algorithms, the goal of aligning the tool zero point with the defect center and the tool axis with the normal vector is transformed into the rotation angles and displacements of each robot joint, thereby driving the robot's movement and completing the posture adjustment. In actual operation, the robot gradually adjusts its position and angle until the ideal welding posture is achieved.

[0064] During the robot's posture adjustment and preparation for welding repair, vision sensors play a crucial monitoring role. The vision sensors continuously acquire image information of the tool and the defect, extracting features such as edge contours and positions of the tool and defect through image processing technology. Based on these features, the real-time relative position data of the tool and defect is calculated and compared with a preset ideal alignment state to determine whether the alignment meets the requirements.

[0065] When the vision sensor detects that the deviation between the tool and the defect exceeds a preset threshold, it indicates that there is an error in the robot's posture or position, which may affect the quality of the weld repair. At this time, the system will trigger a dynamic secondary compensation mechanism. This preset threshold is set according to the accuracy requirements of the weld repair process, and different weld repair scenarios and accuracy requirements will have different thresholds.

[0066] First, a deviation matrix is ​​calculated using real-time relative position data acquired by a vision sensor. This matrix contains information about the tool's deviation from its ideal position in various directions. Next, an iterative nearest-point algorithm is used to match the actual point cloud (i.e., the actual position information of the tool and defect acquired by the vision sensor) with the theoretical model (the ideal alignment model). Based on the matching result, a compensated robot target pose is generated. This target pose corrects the current inaccurate posture, enabling the robot to return to the ideal welding repair position.

[0067] Based on the generated compensated target pose, the robot's motion path is updated. The updated path guides the robot to readjust its posture, moving closer to the ideal alignment. Then, the related operations from acquiring measurement points (similar to step S1) to determining the defect center location and normal vector (similar to step S3) are re-executed, and the alignment status is monitored again by the vision sensor. This cycle continues until the alignment deviation meets the preset accuracy requirements, ensuring that the robot achieves a precise alignment state before welding.

[0068] Based on the foregoing technical solutions, this invention also provides some more specific technical solutions, which are described below.

[0069] In an optional embodiment, the preprocessing in step S2 may further include: The Kalman filter algorithm is used to remove noise from the initial point cloud data; Based on the real-time temperature data of the tank material, the coefficient of linear expansion is dynamically adjusted, and the temperature drift compensation is calculated.

[0070] In some optional implementations, the step of generating the robot obstacle avoidance path in step S4 may include: The robot's joint motion trajectory is planned based on the A* algorithm to avoid obstacles inside the storage tank; A path curvature constraint is introduced to ensure that the deviation between the tool and the defect normal is less than 0.1 mm during the tool movement. The path parameters are transmitted to the robot controller in real time via bus communication, and the risk of joint angle exceeding the limit is monitored.

[0071] During the measurement process, initial point cloud data inevitably contains noise, which may originate from factors such as measurement equipment errors and environmental interference. The Kalman filter algorithm is a highly efficient recursive filter. Based on the system's state equation and observation equation, it uses the estimated value from the previous moment and the observed value from the current moment to iteratively estimate the current state of the system. When processing point cloud data, it can predict the position of the current point based on the statistical characteristics of the data and correct it using actual measurement values, effectively removing noise, preserving the true measurement signal, improving the quality of point cloud data, and providing a more reliable data foundation for subsequent surface reconstruction and defect analysis.

[0072] Because tank materials exhibit thermal expansion and contraction, changes in ambient temperature can alter tank dimensions, leading to coordinate deviations at measurement points and affecting measurement accuracy. By acquiring real-time temperature data of the tank material, the system can dynamically adjust the coefficient of linear expansion. The coefficient of linear expansion reflects the degree of expansion or contraction of a material with temperature changes; different materials have different coefficients of linear expansion at different temperatures. Based on the thermal expansion formula, combined with real-time temperature changes and the dynamically adjusted coefficient of linear expansion, the displacement of measurement points caused by temperature variations—i.e., temperature drift compensation—can be accurately calculated. By compensating for temperature drift in the measurement point coordinates, the influence of temperature on the measurement results can be eliminated, ensuring that the measurement data accurately reflects the actual shape and defect locations of the tank.

[0073] The A* algorithm is a commonly used heuristic search algorithm widely applied in robot path planning. When generating a robot's obstacle avoidance path, the A* algorithm uses the robot's current position as the starting point and the defect center as the target point, while also considering the internal structure of the tank as obstacle information. The algorithm calculates the sum of the actual cost from each node to the starting point and the estimated cost to the target point (i.e., the heuristic function value), and selects the node with the minimum cost for further searching. In this way, among many possible paths, the A* algorithm can quickly find a near-optimal path from the robot's current position, bypassing the internal structure of the tank to reach the defect center, ensuring the robot can safely and efficiently move to the welding repair position in the complex tank environment.

[0074] To ensure the quality of the repair weld, the deviation between the tool and the defect normal must be less than 0.1mm during the tool's movement. Introducing a path curvature constraint limits the curvature of the robot's motion path during path planning. Path curvature reflects the degree of path bending; by setting a reasonable upper limit for curvature, the robot's trajectory is prevented from becoming excessively tortuous, thus ensuring that the tool maintains a small deviation from the defect normal when approaching the defect and performing the repair weld. This helps improve the accuracy of the repair weld, making the weld quality more reliable and preventing weak welds or other quality problems caused by excessive deviation between the tool and the defect normal.

[0075] Real-time transmission of path parameters to the robot controller via bus communication enables efficient data interaction between the path planning system and the robot execution system. Following a specific communication protocol, bus communication quickly and accurately transmits the planned robot motion path parameters, such as the motion angles and speeds of each joint, to the robot controller. Simultaneously, during robot movement, the system continuously monitors joint angles, determining in real-time whether they exceed safe limits. If a joint angle is detected to be close to or exceed the limit, an alarm is immediately issued and corresponding measures are taken, such as adjusting the path planning or pausing robot movement, to prevent robot malfunction or damage due to excessive joint angles, ensuring the safety and stability of robot operation.

[0076] In an optional embodiment, the dynamic secondary compensation in step S5 includes: The real-time relative position data between the tool and the defect is collected by a vision sensor, and the deviation matrix is ​​calculated. The iterative nearest point algorithm is used to match the actual point cloud with the theoretical model to generate the compensated robot target pose. After updating the motion path, repeat steps S1 to S3 until the alignment deviation meets the preset accuracy requirements.

[0077] The vision sensor plays a crucial role in data acquisition during this process. It continuously observes the real-time status of the tool and the defect, using optical imaging and image processing techniques to accurately identify the edges, contours, and other features of the tool and the defect, thereby acquiring their real-time relative position data. This data includes information on the differences between the tool's position and orientation in space and the defect's ideal state. Through specific mathematical algorithms, this positional difference information is quantified to calculate a deviation matrix. The deviation matrix comprehensively describes the degree of deviation of the tool from its ideal position in various directions (such as translational deviations in the X, Y, and Z axes, and rotational deviations around these axes), providing precise data for subsequent adjustments.

[0078] The Iterative Closest Point (ICP) algorithm is the core algorithm for achieving accurate pose compensation. Based on a calculated deviation matrix, this algorithm matches the actual point cloud data (representing the actual relative position of the tool and defect) collected by the vision sensor with a pre-built theoretical model (an ideal alignment model of the tool and defect). During the matching process, the ICP algorithm iteratively optimizes to find an optimal transformation matrix that maximizes the overlap between the actual point cloud and the theoretical model after translation, rotation, and other transformations. Through multiple iterations, when the matching error between the actual point cloud and the theoretical model reaches a certain convergence condition, the resulting transformation matrix is ​​used to generate the compensated robot target pose. This target pose is a correction of the current inaccurate robot pose and can guide the robot to a posture closer to the ideal welding repair position.

[0079] After obtaining the compensated target pose of the robot, the robot's motion path needs to be updated. This is because the robot needs to adjust its movement according to the new target pose, and the original motion path is no longer applicable. The updated motion path will guide the robot to move from its current position to the target pose. Subsequently, the operations from step S1 (operating the robot to drive the contact probe to perform multi-path scanning along the X, Y, and Z axes of the tank, collecting the three-dimensional coordinate data of the measurement points, and generating an initial three-dimensional contour point cloud) to S3 (using the non-uniform rational B-spline algorithm to reconstruct the surface of the corrected point cloud data, generating a high-precision three-dimensional model of the tank's outer surface, and determining the center position and normal vector of the defect based on this model) are repeated. The purpose of this is to reacquire accurate measurement data and reconfirm the alignment status between the tool and the defect. By continuously repeating this cyclical process, the alignment deviation between the tool and the defect is continuously monitored and adjusted. When the alignment deviation meets the preset accuracy requirements, it indicates that the robot's pose and motion path have reached an ideal state, and high-precision repair welding operations can be performed, thereby ensuring the quality of repair welding and improving the reliability and safety of the product.

[0080] In an optional embodiment, the specific steps of NURBS surface reconstruction in step S3 can be: Extract the feature control points from the corrected point cloud data and assign weight coefficients to each control point; The node vector distribution is adaptively adjusted based on the defect geometry to optimize surface smoothness. The NURBS surface equation is fitted using the least squares method, and the coordinates of the defect center and the normal vector are output.

[0081] In some alternative implementations, step S6 may also be included: After the welding is completed, a laser scanner is used to perform three-dimensional morphological inspection of the repaired area and generate a quality assessment report. If residual gaps or deformation exceeding tolerances are detected, mark the defect location and trigger the rework process.

[0082] The corrected point cloud data contains a wealth of information reflecting the surface shape of the tank, but to construct an accurate surface model, feature control points need to be extracted. These points are typically selected at locations where the surface shape changes significantly, such as areas with large curvature changes or edges, and they effectively represent the surface characteristics. Assigning weight coefficients to each control point adjusts its influence on the surface shape. A larger weight indicates a stronger control over the surface shape; by setting appropriate weights, the reconstructed surface can better match the actual shape of the tank.

[0083] In NURBS surface construction, node vectors determine the partitioning of curves or surfaces in the parameter space. Adaptively adjusting the node vector distribution based on defect geometry is crucial for better fitting of surface details and optimizing smoothness. Near defects, node density can be increased to more accurately describe their shape; conversely, in relatively smooth areas of the surface, the number of nodes is reduced to improve computational efficiency. This adjustment allows NURBS surfaces to achieve good fitting results in different regions, ensuring overall surface smoothness and accuracy.

[0084] By fitting the NURBS surface equation using the least squares method and minimizing the sum of squared errors between the reconstructed surface and the corrected point cloud data, various parameters of the NURBS surface are determined, thus constructing a high-precision 3D model of the tank's external surface. Based on this model, the coordinates of the defect center and the normal vector can be accurately output. Determining the defect center coordinates provides a crucial basis for the robot's positioning during subsequent welding; while the normal vector determines the correct orientation of the welding tool, ensuring that the welding operation is perpendicular to the plane where the defect is located, thereby improving the welding quality.

[0085] After the welding repair is completed, a laser scanner is used to perform three-dimensional topographic inspection of the repaired area. The laser scanner acquires high-precision three-dimensional coordinate data of the repaired area by emitting a laser beam and measuring the time or phase change of the reflected light, thereby reconstructing the three-dimensional topography of the repaired area. The reconstructed three-dimensional topography is compared and analyzed with the ideal repair model to evaluate the welding quality from multiple aspects, such as the flatness of the welded area, the transition with the surrounding area, and the thickness of the weld material buildup. A detailed quality assessment report is generated, comprehensively reflecting the effect of the welding repair.

[0086] If residual gaps or deformation exceeding tolerances are detected in the quality assessment report, it indicates that the repair welding quality has not met the requirements. In this case, the system will mark the defect location and record the specific coordinates of the non-compliant area for subsequent investigation and processing. Simultaneously, a rework process will be triggered, involving re-measurement, repositioning, and re-welding until the quality of the repaired area meets preset standards, ensuring the safety and reliability of aerospace tanks and other equipment.

[0087] For circular defects, their geometric characteristics, such as dimensions (diameter, etc.) and form and position errors (roundness, cylindricity, positional accuracy, etc.), cannot be directly measured. Instead, they are obtained through indirect measurement. This involves transforming the measurement of a geometric element into the measurement of the coordinates of a specific set of points on that element. After obtaining the coordinate values ​​of these points, appropriate mathematical calculations can be used to deduce the dimensions and form and position errors of the defect, thus providing a comprehensive and accurate quantitative description of the defect.

[0088] Figure 2 This diagram illustrates the method for measuring circular holes on the outer surface of a storage tank, using a circular defect as an example. When measuring the diameter of a circular hole on the tank's outer surface, the measurement is performed within section I, perpendicular to the hole's axis. This is because the geometric features of a circle can be presented in two dimensions on this section, facilitating measurement and calculation. Three points are touched on the inner wall of section I, labeled as point 1, point 2, and point 3. Mathematically, three points not on the same straight line can define a unique circle. Assume the coordinates of these three points are... , , According to the standard equation of a circle (in Let the coordinates be the center of the circle. Let be the radius. Substituting the coordinates of the three points into the equation, we obtain a system of three quadratic equations in three variables: By solving this system of equations, the coordinates of the circle's center can be obtained. The diameter d = 2r. In practical calculations, the calculation process can be simplified using some mathematical techniques. For example, the distances between each pair of the three points can be calculated first, and the geometric properties of the triangle can be used to help solve for the coordinates of the center and the radius.

[0089] When measuring the roundness error of a cross-section, measuring only three points is insufficient; more points within the cross-section need to be measured. The more points measured, the more accurate the roundness measurement. In this case, the roundness error can be calculated using the least squares method or the minimum condition method. The principle of the least squares method is to determine the best-fit circle by minimizing the sum of the squares of the distances from all measurement points to the fitted circle. Let the coordinates of the measurement points be... The equation of the fitted circle is Then the objective function is Through the , , Find the partial derivatives and set them equal to 0, then solve the system of equations to obtain the parameters of the fitted circle. The minimum condition method finds two concentric circles that encompass all measurement points, based on the principle of minimum containment area; the difference in radii between these two concentric circles is the roundness error. This method better matches the definition of roundness error, but the calculation process is relatively complex and usually requires iterative algorithms. In practical applications, the appropriate method must be selected based on the required measurement accuracy and computational resource limitations.

[0090] To measure the cylindricity error of the defect and determine its axis position, multiple cross-sectional circles (I, II, ..., m, where m is the number of the measured cross-sectional circles) perpendicular to the defect axis need to be measured. At each cross-section, the center coordinates are obtained using the method of measuring the diameter and center coordinates. , … The calculation of cylindricity error can also be based on the least squares method or the minimum condition method. With the least squares method, a spatial fit is first performed on all measurement points to obtain a fitted cylindrical surface. Then, the distance from each measurement point to this fitted cylindrical surface is calculated, and the cylindricity error is determined by statistical analysis of these distances. With the minimum condition method, the smallest cylindrical surface that encompasses all measurement points is found; the difference in radius of this smallest cylindrical surface is the cylindricity error. When determining the location of the defect axis, since the coordinates of the centers of multiple cross-sectional circles are already obtained, these centers should ideally lie on the same straight line, i.e., on the defect axis. A spatial straight line equation can be obtained by fitting these center coordinates; this straight line represents the defect axis. A commonly used fitting method is least squares linear fitting, which determines the parameters of the straight line by minimizing the sum of the squares of the perpendicular distances from each circle center to the fitted straight line.

[0091] Three points are touched on the defective end face A. Assume the coordinates of these three points are... , , First, based on these three points, fit the plane equation of end face A. The general form of the plane equation is: By substituting the coordinates of the three points into the equation, a system of equations for A, B, C, and D is obtained. Solving this system of equations determines the parameters of the plane equation. After determining the plane equation of end face A, the intersection point of the axis and the end face is calculated by combining it with the previously obtained defect axis equation. By comparing the deviation between the actual position and the theoretical position (usually the ideal position determined according to design requirements) of the intersection point, the positional error of the defect axis relative to the end face can be calculated. The positional error comprehensively reflects the degree of spatial offset of the axis relative to the end face and is one of the important indicators for evaluating the impact of defects on the overall performance of the equipment.

[0092] It should be noted that the various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. The present invention is not limited to the specific structures described above and shown in the figures. Furthermore, for the sake of brevity, detailed descriptions of known methods and techniques are omitted here.

[0093] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A coordinate measuring and positioning system for welding pull-out friction plugs, characterized in that, include: The robot body has a spindle tool holder at its end. A coordinate measuring sensor is mounted on the spindle tool holder via a clamping adapter, and is used to contact the surface of the tank and collect the coordinates of the measurement points. The control subsystem is used to latch the grating signal of the measurement sensor, record the three-dimensional coordinates of the measurement point and generate point cloud data. The data processing module is used to perform surface fitting on the point cloud data, reconstruct the tank surface model, and calculate the center position and normal of the defect; The robot controller is used to adjust the robot's posture according to the center position and normal of the defect so that the tool zero point coincides with the center of the defect and the cutting tool coincides with the normal.

2. The coordinate measuring and positioning system for pull-out friction plug repair welding according to claim 1, characterized in that, The clamping adapter includes: The tool holder connection part is detachably connected to the spindle tool holder; The probe mounting part is used to fix the coordinate measuring sensor, and a radius compensation adjustment structure is provided between the probe mounting part and the tool holder connection part to compensate for the influence of the probe radius on the measurement accuracy.

3. A coordinate measuring and positioning method for welding pull-out friction plugs, applied to the coordinate measuring and positioning system for welding pull-out friction plugs as described in any one of claims 1 to 2, characterized in that, Includes the following steps: S1. The robot drives the contact probe to perform multi-path scanning along the X, Y, and Z axes of the storage tank, collects the three-dimensional coordinate data of the measurement points, and generates an initial three-dimensional contour point cloud. S2. Preprocess the initial three-dimensional contour point cloud, including contact point offset compensation based on probe radius and temperature drift compensation based on material linear expansion coefficient, to obtain corrected point cloud data. S3. The non-uniform rational B-spline algorithm is used to reconstruct the surface of the corrected point cloud data to generate a high-precision three-dimensional model of the tank's outer surface, and the center position and normal vector of the defect are determined based on the model. S4. The defect center coordinates and normal vector are transmitted to the robot control system via bus communication. Combined with the defect edge data scanned synchronously by the laser rangefinder, the robot obstacle avoidance motion path is generated. S5: Adjust the robot's posture to make the tool zero point coincide with the defect center and the tool axis aligned with the normal vector, and monitor the alignment status of the tool and the defect in real time through a vision sensor; if the deviation exceeds the preset threshold, trigger dynamic secondary compensation and update the robot's motion path.

4. The coordinate measuring and positioning method for pull-out friction plug repair welding according to claim 3, characterized in that, The preprocessing in step S2 also includes: The Kalman filter algorithm is used to remove noise from the initial point cloud data; Based on the real-time temperature data of the tank material, the coefficient of linear expansion is dynamically adjusted, and the temperature drift compensation is calculated.

5. The coordinate measuring and positioning method for pull-out friction plug repair welding according to claim 3, characterized in that, The generation of the robot obstacle avoidance path in step S4 includes: The robot's joint motion trajectory is planned based on the A* algorithm to avoid obstacles inside the storage tank; A path curvature constraint is introduced to ensure that the deviation between the tool and the defect normal is less than 0.1 mm during the tool movement. The path parameters are transmitted to the robot controller in real time via bus communication, and the risk of joint angle exceeding the limit is monitored.

6. The coordinate measuring and positioning method for pull-out friction plug repair welding according to claim 3, characterized in that, The dynamic secondary compensation in step S5 includes: The real-time relative position data between the tool and the defect is collected by a vision sensor, and the deviation matrix is ​​calculated. The iterative nearest point algorithm is used to match the actual point cloud with the theoretical model to generate the compensated robot target pose. After updating the motion path, repeat steps S1 to S3 until the alignment deviation meets the preset accuracy requirements.

7. The coordinate measuring and positioning method for pull-out friction plug repair welding according to claim 3, characterized in that, The specific steps of NURBS surface reconstruction in step S3 are as follows: Extract the feature control points from the corrected point cloud data and assign weight coefficients to each control point; The node vector distribution is adaptively adjusted based on the defect geometry to optimize surface smoothness. The NURBS surface equation is fitted using the least squares method, and the coordinates of the defect center and the normal vector are output.

8. The coordinate measuring and positioning method for pull-out friction plug repair welding according to claim 3, characterized in that, It also includes step S6: After the welding is completed, a laser scanner is used to perform three-dimensional morphological inspection of the repaired area and generate a quality assessment report. If residual gaps or deformation exceeding tolerances are detected, mark the defect location and trigger the rework process.

Images